3D FINITE-ELEMENT CODE APPLICATION IN HOT-ROLLING

The aim of our paper is to show the contribution of 3D finite element modelling in the field of metal flow prevision in hot rolling. Improving the quality of rolled products implies a more and more refined thermomechanical analysis of the rolling process. Due to increasing computing power, 3D modelling of industrial cases with finite element codes becomes easier. The approach used here has been to develop an existing code, FORGE3, made by CEMEF (Ecole des Mines de Paris), to allow hot rolling modelling. FORGE3 mechanical formulation The main features may be summarized : - non-steady thermomechanical formulation, expressed in terms of velocity; - viscoplastic Norton-Hoff behaviour with sensitivity to temperature and equivalent strain; - Norton friction law; Concerning boundary conditions: - rigid non-steady tools for forging; - steady tools for rolling. Classical approach in rolling modelling consists in Eulerian analysis. The non-steady approach used here is CPU time consuming but it allows more general situations. Model validation (round/oval rolling) During round/oval pass, equivalent strain distribution in the product may be non-uniform, depending on rolling conditions. In order to validate the code, we compute an experimental round-oval pass made in British Steel Laboratory (ECSC 7210 EB/805). Computed strain values are compared to experimental measurement, showing quite good agreement. Application example (sizing-press). This example shows a comparison between two slab sizing processes, in the same simplified conditions: vertical rolling and sizing press. The sizing press process is supposed to allow greater width reductions with less geometrical problems. Despite of his non-steady nature, this process has a lot of common features with rolling (neutral point flow...). In order to approach real situations, we have computed the following problems, with simplifying assumptions : - 200 mm width reduction with sizing-press followed by 20 mm horizontal rolling reduction ; - 200 mm vertical rolling width reduction followed by 20 mm horizontal rolling reduction; - 2 x 100 mm vertical rolling width reduction followed by horizontal rolling giving the same final thickness. Concerning thickness and width deviations, the sizing press shows better performance, giving higher efficiency after horizontal rolling (about 20 %). Computed slab profiles are in good agreement with other published results (SMS). Conclusion Comparisons between computed and measured strain values on round-oval pass have been made to validate a finite element code, FORGE3. In this paper, this code was used as an investigation tool to assess the mechanical superiority of the sizing press process on rolling, for heavy slab reductions.